This week in AI: trust just got a price tag

"People will trust AI more than they trust each other." That's what Yuval Noah Harari said in a recent interview. It sounded like a prediction. This week, two other stories made it look more like the present. Trust is turning into something you pay for Harari's argument is that AI is more honest and makes fewer mistakes than people do, so people will start leaning on it instead of leaning on each other. You'd think humans would be better judges of what's real. Turns out the opposite. Gartner estimates the world will spend 25 billion dollars by 2030 just verifying things: whether a message is real, a call is real, a transaction is real. That's not cheap. Next time you get a call and can't tell if it's a real person, the verification step alone might take longer than you'd expect. Skill can be borrowed too, and it comes with interest A study in the Proceedings of the National Academy of Sciences had students use AI to help solve problems. Scores went up 48 percent. Sounds great. Here's the catch: take the AI away and test them again, and scores dropped 17 percent, worse than students who never used AI at all. The researchers tried another approach, having experts design a version that guided students through the problem instead of letting them skip to the answer. That version pushed the gain to 127 percent. If you use AI to check your kid's homework, cram for a test, or pick up a new skill, you might just be borrowing ability you'll have to pay back later. This week, that same instinct to fake it turned up somewhere even more absurd: resumes. People are rewriting their old job titles too A Stanford study found nearly one in five American LinkedIn users go back and rewrite job descriptions from positions they left years ago, stuffing today's AI buzzwords into roles that existed before AI did. MBAs do it the most. Almost 30 percent of them. In tech, the title "AI product manager" gets edited a staggering 71 percent of the time. The same survey asked employers too: only one in ten workers think job postings are fully honest. Job seekers are performing. Companies writing the listings are performing. They meet on the page, and neither one believes the other. If resumes and job postings are both theater, how do you tell who's actually on the other side. Maybe Harari's right after all. Maybe you'll end up asking AI instead.